Paul, 

Below pls find the output of gdb. Are the symbols output-ed adequate? I'll 
check into the -d status of our build... 


On dtrace: A _very_ powerful tool; allows us to monitor 'probes' placed in 
running processes. 
Keyword, though, is 'running'; this may not help us here... 


Lou 


# gdb /usr/local/sbin/dbmail-lmtpd 
GNU gdb 6.6 
(...) 
This GDB was configured as "sparc-sun-solaris2.10"... 


(gdb) run -D 
Starting program: /usr/local/sbin/dbmail-lmtpd -D 
[New LWP 1] 
[New LWP 2] 
[LWP 2 exited] 
[New LWP 2] 
warning: Temporarily disabling breakpoints for unloaded shared library 
"/usr/lib/ld.so.1" 
warning: Lowest section in /lib/libpthread.so.1 is .dynamic at 00000074 
warning: Lowest section in /lib/libthread.so.1 is .dynamic at 00000074 
warning: Lowest section in /lib/libdl.so.1 is .hash at 000000b4 
[LWP 2 exited] 
[New LWP 2] 
Aug 02 07:57:47 leeloo [8925]: Debug:[serverparent] 
serverparent.c,LoadServerConfig(+267): server will create [2] children 
Aug 02 07:57:47 leeloo [8925]: Debug:[serverparent] 
serverparent.c,LoadServerConfig(+280): children will make max. [10000] 
connections 
Aug 02 07:57:47 leeloo [8925]: Debug:[serverparent] 
serverparent.c,LoadServerConfig(+293): timeout [300] seconds 
Aug 02 07:57:47 leeloo [8925]: Debug:[serverparent] 
serverparent.c,LoadServerConfig(+305): login_timeout [60] seconds 
Aug 02 07:57:47 leeloo [8925]: Debug:[serverparent] 
serverparent.c,LoadServerConfig(+310): no value for SOCKET in config file 
Aug 02 07:57:47 leeloo [8925]: Debug:[serverparent] 
serverparent.c,LoadServerConfig(+313): socket [] 
Aug 02 07:57:47 leeloo [8925]: Debug:[serverparent] 
serverparent.c,LoadServerConfig(+325): binding to PORT [24] 
Aug 02 07:57:47 leeloo [8925]: Debug:[serverparent] 
serverparent.c,LoadServerConfig(+346): binding to IP [*] 
Aug 02 07:57:47 leeloo [8925]: Debug:[serverparent] 
serverparent.c,LoadServerConfig(+353): no value for BACKLOG in config file. 
Using default value [16] 
Aug 02 07:57:47 leeloo [8925]: Debug:[serverparent] 
serverparent.c,LoadServerConfig(+367): resolving client IP 
Aug 02 07:57:47 leeloo [8925]: Debug:[serverparent] 
serverparent.c,LoadServerConfig(+381): Enabling LMTP-before-SMTP 
Aug 02 07:57:47 leeloo [8925]: Debug:[serverparent] 
serverparent.c,LoadServerConfig(+393): effective user shall be [nobody] 
Aug 02 07:57:47 leeloo [8925]: Debug:[serverparent] 
serverparent.c,LoadServerConfig(+405): effective group shall be [nogroup] 
Aug 02 07:57:47 leeloo [8925]: Debug:[serverparent] 
serverparent.c,LoadServerConfig(+417): will maintain minimum of [2] spare 
children in reserve 
Aug 02 07:57:47 leeloo [8925]: Debug:[serverparent] 
serverparent.c,LoadServerConfig(+429): will maintain maximum of [4] spare 
children in reserve 
Aug 02 07:57:47 leeloo [8925]: Debug:[serverparent] 
serverparent.c,LoadServerConfig(+441): will allow maximum of [10] children 
Aug 02 07:57:47 leeloo [8925]: Debug:[server] pool.c,statefile_create(+666): 
Creating scoreboard at [/var/run/dbmail-lmtpd.state]. 


Program received signal SIGSEGV, Segmentation fault. 
CreateSocket (conf=0xffbfe0d0) at server.c:460 
460 if ((sock = socket(res->ai_family, res->ai_socktype, res->ai_protocol)) < 
0) { 


(gdb) bt 
#0 CreateSocket (conf=0xffbfe0d0) at server.c:460 
#1 0xff35fd60 in server_run (conf=0xffbfe0d0) at server.c:263 
#2 0xff3623dc in serverparent_mainloop (config=0xffbfe0d0, service=0x14680 
"LMTP", servicename=0x14688 "dbmail-lmtpd") at serverparent.c:182 
#3 0x00012368 in main (argc=83592, argv=0xffbffd94) at lmtpd.c:55 



----- Original Message ----- 
From: "Paul J Stevens" <[email protected]> 
To: "DBMail mailinglist" <[email protected]> 
Sent: Monday, August 2, 2010 5:06:03 AM 
Subject: Re: [Dbmail] DBMail daemons segfaulting 



try: 

gdb /usr/local/sbin/dbmail-lmtpd 
(gdb) run -D 

that should segfault. When it does do: 

(gdb) bt 

and send the output. 

All this requires you to build dbmail with debugging symbols (-g) and 
without stripping the binaries post-build. 

I've never worked with dtrace. No idea what it does or what it's 
capabilities are. 




On 08/02/2010 03:47 AM, Lou Picciano wrote: 
> Paul, 
> 
> Please specify exactly what tests you want performed with gdb - happy to 
> help... 
> Where would you like the backtrace from? 
> 
> Remember, it's Solaris - we also have dtrace at our disposal. 
> 
> Here's the output of a run command from gdb: 
> 
> (gdb) file /usr/local/sbin/dbmail-lmtpd 
> Reading symbols from /usr/local/sbin/dbmail-lmtpd...done. 
> (gdb) run 
> Starting program: /usr/local/sbin/dbmail-lmtpd file dbmail-lmtpd 
> [New LWP 1] 
> [New LWP 2] 
> [LWP 2 exited] 
> [New LWP 2] 
> warning: Temporarily disabling breakpoints for unloaded shared library 
> "/usr/lib/ld.so.1" 
> warning: Lowest section in /lib/libpthread.so.1 is .dynamic at 00000074 
> warning: Lowest section in /lib/libthread.so.1 is .dynamic at 00000074 
> warning: Lowest section in /lib/libdl.so.1 is .hash at 000000b4 
> /usr/local/sbin/dbmail-lmtpd: unrecognized options: file dbmail-lmtpd 
> 
> *** dbmail-lmtpd *** 
> This daemon provides Local Mail Transport Protocol services. 
> See the man page for more info. 
> 
> Common options for all DBMail daemons: 
> (...) 
> 
> Program exited with code 01. 
> -------------------- 
> 
> Also, when run non-daemonized, it simply outputs the same log output 
> (sent earlier), then quits back to shell. No message of segfault. And 
> it's not visible in the process list at that point, either. 
> 
> Pls tell me next steps. Lou 
> 
> ----- Original Message ----- 
> From: "Paul J Stevens" <[email protected]> 
> To: "DBMail mailinglist" <[email protected]> 
> Sent: Sunday, August 1, 2010 5:38:17 PM 
> Subject: Re: [Dbmail] DBMail daemons segfaulting 
> 
> On 08/01/2010 08:12 PM, Lou Picciano wrote: 
> ... 
> 
>> As to version we are building: 
>> 
>> # git branch -v 
>> * dbmail_2_2 793da8e massive speedup in imap-append 
> 
> Ok. Just fine and dandy. 
> 
> 
> 
>> Here is the entire output of the dbmail-lmtpd command, including segfault: 
>> 
>> # dbmail-lmtpd -D 
> 
>> pool...c,statefile_create(+666): Creating scoreboard at 
>> [/var/run/dbmail-lmtpd.state]. 
>> Segmentation Fault (core dumped) 
> 
> Not good! Is that what happens when you run the normal daemonizing mode? 
> 
>> 
>> This run does write the following to /var/log/dbmail.err (but we have 
>> /_never_/ seen any output to dbmail.log): 
>> 
>> Aug 01 12:01:46 leeloo [3170]: Debug:[server] 
>> server.c,server_daemonize(+238): sid: [3169] 
>> Aug 01 12:01:46 leeloo [3170]: Debug:[server] 
>> pool.c,statefile_create(+666): Creating scoreboard at 
>> [/var/run/dbmail-lmtpd.state]. 
>> 
>> We do have ALL of the debuggers on in the dbmail and ALL of the daemon 
>> sections of the dbmail.conf file. 
>> 
>> TRACE_LEVEL = 5 
>> TRACE_STDERR = 5 
>> TRACE_SYSLOG = 5 
> 
> All generate the same output. The first is deprecated, equals the third 
> option. The second appends the log lines to the error log, the third 
> option sends it to the syslog facility. 
> 
> Can you run the command through gdb? (That's what the -D switch is for). 
> A backtrace would really be helpful. 
> 
> 
> -- 
> ________________________________________________________________ 
> Paul Stevens paul at nfg.nl 
> NET FACILITIES GROUP GPG/PGP: 1024D/11F8CD31 
> The Netherlands________________________________http://www.nfg.nl 
> _______________________________________________ 
> DBmail mailing list 
> [email protected] 
> http://mailman.fastxs.nl/cgi-bin/mailman/listinfo/dbmail 
> 
> 
> 
> _______________________________________________ 
> DBmail mailing list 
> [email protected] 
> http://mailman.fastxs.nl/cgi-bin/mailman/listinfo/dbmail 


-- 
________________________________________________________________ 
Paul Stevens paul at nfg.nl 
NET FACILITIES GROUP GPG/PGP: 1024D/11F8CD31 
The Netherlands________________________________http://www.nfg.nl 
_______________________________________________ 
DBmail mailing list 
[email protected] 
http://mailman.fastxs.nl/cgi-bin/mailman/listinfo/dbmail 
_______________________________________________
DBmail mailing list
[email protected]
http://mailman.fastxs.nl/cgi-bin/mailman/listinfo/dbmail

Reply via email to